Research On Identification Technology Based On Hand Shape And Palmprint

There is no possibility to achieve the perfect identification system based on the single hand biometric such as hand shape and palmprint.The new research direction to multi-model biometric fusion is dual-model biometric technology which based on hand shape and palmprint.This paper focuses on the dual-modal identification method of decision-making layer hand shape and palmprint identification system which based on the analysis of hand shape and palmprint two features identification technology,according to the characteristic between the hand image acquisition can get hand shape characteristic and palmprint feature,the use of hand shape matching speed and palmprint recognition rate advantage,to meet the needs of the application of high precision and speed requirements.This research can provide scientific basis and technical method for the construction of embedded hand features identification system,containing certain scientific value and application value.The main research contents of this paper are as follows:First of all,using hand shape and palmprint as hand characteristics,can extract from the hand image acquisition one-time,this paper respectively uses the camera and cameras capturing the hand images of 50 students,built the corresponding hand image library,reduced the complexity of capture.Secondly,according to the study,through extraction hand shape and palmprint feature from a hand image pretreatment,the original image(640×480)gray,gauss filtering,binaryzation,morphological process,contour extraction and feature point position can use a treatment process,effective sharing method in the hand shape and palmprint image,can reduce the processing time of pretreatment of multi feature fusion.In addition,in the research of feature point positioning,this paper present a scan line positioning point fingers and hand shape apex defect positioning root point method,which can position hand shape and palmprint key point one-time,the experimental results show that this method is better than the linear fitting method of locating feature point.Third,according to the location of feature points,this paper extracts hand shape features using feature vector method;extracted palmprint features of hand image is needed by rotation correction,extraction of region of interest image(ROI 128×128),down sampling and histogram equalization,and then uses the PCA algorithm.Fourth,this paper uses the nearest neighbor classification method,and is based on analyzing the hand shape recognition and palmprint recognition,which use decision fusion theory to analysis hand shape and palmprint series parallel fusion methods.It focus on three types of comparison based on hand shape and palmprint series method,and put forward series by the user verification,and this method can adapt to all the training samples better,and the system becomes more robust.Finally,on the basis of theoretical research,this paper used VC++6.0 as the development platform,and opencv as the auxiliary tool,completed the development of identity recognition system based on hand shape,palmprint and a variety fusion,and it respectively to test image library with camera and cameras acquisition.It can be found that the method is based on hand shape and palmprint series and parallel recognition is better than the single hand shape recognition or palmprint recognition in recognition rate by comparing the hand shape recognition,palmprint recognition,based on hand shape and palmprint series and parallel fusion recognition results,and the user authentication series recognition rate can reach 97%,the recognition time is 78ms.